Cone Beam Computed Tomography (CBCT) is a widely adopted technology in dental healthcare, particularly for surgical planning. However, effective visualization methods are crucial for formatting and analyzing volumetri...
详细信息
ISBN:
(数字)9798350373271
ISBN:
(纸本)9798350373288
Cone Beam Computed Tomography (CBCT) is a widely adopted technology in dental healthcare, particularly for surgical planning. However, effective visualization methods are crucial for formatting and analyzing volumetric data. Two primary formatting options - Curved Multi-planar Reformatting and Direct Volume Rendering - are used widely for evaluating jaw shape and creating panoramic views. Existing software methods frequently require the manual selection of several nodes along the jaw on an axial image that roughly represents the shape of the dental arch. Our work intends to introduce a novel and automatic algorithm for detecting the patient’s jaw from CBCT, suitable for defining a focal trough and evaluating an optimal patient-specific dynamic rotation trajectory applicable to rotational panoramic radiography based only on the CBCT volume data without the need to delineate a dental arch. Our findings showed that this approach works well even for patients without teeth or having heavy metal implants.
Accurate and timely diagnosis of liver disorders such as fatty liver disease, chronic viral hepatitis, and excessive alcohol consumption is crucial for maintaining liver health. Traditional methods for liver screening...
详细信息
ISBN:
(数字)9798331531539
ISBN:
(纸本)9798331531546
Accurate and timely diagnosis of liver disorders such as fatty liver disease, chronic viral hepatitis, and excessive alcohol consumption is crucial for maintaining liver health. Traditional methods for liver screening are often subjective, time-consuming, and reliant on the expertise of the sonographer, which can impact diagnostic accuracy. To tackle these challenges, in this paper proposed a deep learning (DL) based framework to enhance the effective diagnosis of focal liver lesions. This approach leverages an channel attention mechanism integrated with the DarkNet-19 pre-trained model to improve feature extraction and boosts classification accuracy. By automating the diagnostic process, the proposed model addresses the limitations of traditional methods, providing a more efficient and reliable solution for liver disorder diagnosis. Experimental results with an ultrasound image dataset demonstrate that the proposed model significantly outperforms conventional DL methods, showcasing its advantages in performance and efficiency.
Crowd anomaly detection suffers from limited training data under weak supervision. In this paper, we propose a dual-mode iterative denoiser to tackle the weak label challenge for anomaly detection. First, we use a con...
详细信息
Crowd anomaly detection suffers from limited training data under weak supervision. In this paper, we propose a dual-mode iterative denoiser to tackle the weak label challenge for anomaly detection. First, we use a convolution autoencoder (CAE) in image space to act as a cluster for grouping similar video clips, where the spatial-temporal similarity helps the cluster metric to represent the reconstruction error. Then we use the graph convolution neural network (GCN) to explore the temporal correlation and the feature similarity between video clips within different rough labels, where the classifier can be constantly updated in the label denoising process. Without specific image-level labels, our model can predict the clip-level anomaly probabilities for videos. Extensive experiment results on two public datasets show that our approach performs favorably against the state-of-the-art methods.
Among various types of skin cancers, melanoma is the most aggressive and deadly. There is a notable growth in the implementation of deep learning (DL) methods to identify skin malignancies in dermoscopy images. This p...
详细信息
ISBN:
(数字)9798350350845
ISBN:
(纸本)9798350350852
Among various types of skin cancers, melanoma is the most aggressive and deadly. There is a notable growth in the implementation of deep learning (DL) methods to identify skin malignancies in dermoscopy images. This paper introduces a lightweight DL-based approach designed for seamless integration into low-memory devices within healthcare applications. The proposed method incorporates three lightweight convolutional neural network (CNN) models: MobileNet-v2, SqueezeNet, and GoogLeNet. Initially, test features are computed from fine-tuned deep CNN models. Subsequently, probability scores for each class are derived by training and testing a random forest classifier with features extracted from the models. Then, the proposed method uses an average ensemble voting technique on the probability scores to enhance the classification performance compared to the individual models. The proposed of lightweight CNN model demonstrated an accuracy of 85.19 % which is better than existing works.
Breast cancer (BC) is a potentially life-threatening disease that occurs because of uncontrolled growth of abnormal corpuscles in the breast tissue. Pathologists analyze the tissue structures using histopathological w...
详细信息
ISBN:
(数字)9798350350951
ISBN:
(纸本)9798350350968
Breast cancer (BC) is a potentially life-threatening disease that occurs because of uncontrolled growth of abnormal corpuscles in the breast tissue. Pathologists analyze the tissue structures using histopathological whole slide images to identify cancerous anomalies. However, pathologists face severe challenges such as fatigue, subjectivity, and inter-observer variability in the early detection of BC. Understanding the intricacies of BC from molecular tissue structures is complex, and inexpertise leads to adverse outcomes. This paper proposes a computed aided detection (CAD) system that can assist histopathologists in the early detection of BC, potentially reducing the abnormalities and diagnostic time. Leveraging the power of convolutional neural networks (CNNs), a stacked ensemble-based model is developed to identify benign and malignant cancerous tissues using histopathological images. The ensemble models comprise three deep CNNs, namely MobileNetV2, ShuffleNet, and SqueezeNet, trained on the BreakHis dataset. Finally, individual CNNs predictions are fed to the average voting-based classifier to identify benign and malignant tissues. The stacked ensemble-based deep CNN model outperformed the individual CNN models in BC prediction, achieving superior accuracy and robustness.
Traditional video synopsis methods model the processing into an optimization formula where relations among objects such as collision cost are utilized while entire re-calculation is introduced under each possible temp...
详细信息
ISBN:
(纸本)9781509028610
Traditional video synopsis methods model the processing into an optimization formula where relations among objects such as collision cost are utilized while entire re-calculation is introduced under each possible temporal shift. Unlike the pairwise cost optimization, we propose a low-complexity and efficient online synopsis method where each tube is processed independently. Without tubes' comparison, the rearrangement is accomplished by a simple projection strategy and an updating projection matrix which records the newest information of the moving space. Furthermore, buffer and a predefined fitness condition also help to increase spatial and temporal utilization. Experiment results demonstrate that the proposed method is superior to other synopsis methods in the processing speed and temporal consistency.
Since the lighting conditions in strong contrast regions between the light and dark cant be estimated accurately by traditional center/surround Retinex algorithm, the over-enhancement and color distortion may exist. I...
详细信息
ISBN:
(纸本)9781509028610
Since the lighting conditions in strong contrast regions between the light and dark cant be estimated accurately by traditional center/surround Retinex algorithm, the over-enhancement and color distortion may exist. In view of this, combining with the human visual characteristics, a color image enhancement algorithm based on tone-preserving was proposed. A determination function was added to the bilateral filter to estimate illuminance image more accurately and weaken over-enhancement. According to human visual masking effect, the improved gamma correction was utilized to correct the brightness of illumination image adaptively and the local contrast of reflection image obtained by division was enhanced based on local statistics. Besides, the final enhanced image was obtained by combining illumination image with reflection image, which can make image appear more natural. Compared with other similar algorithms from both subjective and objective aspects, the results show that this method being applied to low-contrast color image enhancement can not only improve image clarity, but reduce color distortion.
作者：
王振华吴伟仁田玉龙田金文柳健Institute for Pattern Recognition and Artificial Intelligence
State Key Lab for Image Processing and Intelligent ControlHuazhong University of Science and Technology Wuhan 430074 China Institute for Pattern Recognition and Artificial Intelligence
State Key Lab for Image Processing and Intelligent ControlHuazhong University of Science and Technology Wuhan 430074 China major limitation for deep space communication is the limited bandwidths available. The downlink rate using X-band with an L2 halo orbit is estimated to be of only 5.35 GB/d. However the Next Generation Space Telescope (NGST) will produce about 600 GB/d. Clearly the volume of data to downlink must be reduced by at least a factor of 100. One of the resolutions is to encode the data using very low bit rate image compression techniques. An very low bit rate image compression method based on region of interest(ROI) has been proposed for deep space image. The conventional image compression algorithms which encode the original data without any data analysis can maintain very good details and haven't high compression rate while the modern image compressions with semantic organization can have high compression rate even to be hundred and can't maintain too much details. The algorithms based on region of interest inheriting from the two previews algorithms have good semantic features and high fidelity and is therefore suitable for applications at a low bit rate. The proposed method extracts the region of interest by texture analysis after wavelet transform and gains optimal local quality with bit rate control. The Result shows that our method can maintain more details in ROI than general image compression algorithm(SPIHT) under the condition of sacrificing the quality of other uninterested areas
A major limitation for deep space communication is the limited bandwidths available. The downlinkrate using X-band with an L2 halo orbit is estimated to be of only 5.35 GB/d. However, the Next GenerationSpace Telescop...
详细信息
A major limitation for deep space communication is the limited bandwidths available. The downlinkrate using X-band with an L2 halo orbit is estimated to be of only 5.35 GB/d. However, the Next GenerationSpace Telescope (NGST) will produce about 600 GB/d. Clearly the volume of data to downlink must be re-duced by at least a factor of 100. One of the resolutions is to encode the data using very low bit rate image com-pression techniques. An very low bit rate image compression method based on region of interest(ROI) has beenproposed for deep space image. The conventional image compression algorithms which encode the original datawithout any data analysis can maintain very good details and haven' t high compression rate while the modernimage compressions with semantic organization can have high compression rate even to be hundred and can' tmaintain too much details. The algorithms based on region of interest inheriting from the two previews algorithmshave good semantic features and high fidelity, and is therefore suitable for applications at a low bit rate. Theproposed method extracts the region of interest by texture analysis after wavelet transform and gains optimal localquality with bit rate control. The Result shows that our method can maintain more details in ROI than generalimage compression algorithm(SPIHT) under the condition of sacrificing the quality of other uninterested areas.
Target extraction is a key technology for image measurement of moving particles distributed in fluidic system. In this paper, we propose a novel moving particle extraction method based on multimodal characteristic of ...
详细信息
The analysis of large-scale crowd behavior plays a crucial role in public safety. However, intelligent systems face three major challenges in analyzing dense crowd behavior: the severe occlusion between individuals, t...
详细信息
The analysis of large-scale crowd behavior plays a crucial role in public safety. However, intelligent systems face three major challenges in analyzing dense crowd behavior: the severe occlusion between individuals, the variability in behavior patterns, and the complexity of behavioral evolution. To address these challenges, we propose the Physics-Environment Interaction Network (PEIN), which directly models the motion characteristics of a group with its physics attributes. Specifically, our method consists of two streams. The first stream is the physics-informed crowd property stream, which leverages on the similarity between dense crowd motion and fluid dynamics, using the Navier-Stokes (N-S) equation from fluid mechanics as the modeling framework to describe crowd motion. Considering the inherent relationship between the terms in the N-S equation and various crowd properties (collectiveness, conflict, uniformity and stability), we model these terms with operators and neural networks guided by these crowd properties, enabling the modeling of crowd motion characteristic without relying on the extraction of individual motion information. The second stream is the environment perception stream. Considering that the physics-informed crowd property stream mainly focuses on instantaneous information and that scenes with dense crowd behavior have variability, we introduce a 3D network to enhance the model's robustness. This stream can extract global spatiotemporal information from input video frames, providing a comprehensive perception of the surrounding crowd environment. Since the two streams process different data sources, we design a dual cross-attention mechanism to to enable between features from different modalities, resulting in a joint learnable representation for the final crowd behavior recognition. By incorporating physical laws as constraints, we design a physics-informed loss function combined with a crowd behavior loss function to optimize the model. Consi
暂无评论